JP2006049967A - Clock signal method and clock signal extracting apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable extracting a clock signal from which a timing jitter is eliminated even if the large timing jitter is contained in an optical pulse signal. <P>SOLUTION: A clock signal extracting apparatus 125 consists of an optical modulating unit 120 and a clock signal/feedback signal generating unit 100. The optical modulating unit is configured in such a way that a first EAM 92 for modulating an optical pulse signal, an optical amplifier 94 for amplifying an optical output signal 93 of the first EAM 92, and a second EAM 96 for modulating an output 95 of this optical amplifier are connected in cascade in the order of the first EAM, optical amplifier and the second EAM. An optical pulse signal 48 is input into the optical modulating unit, modulated by an electric signal 79 and outputted as a modulated optical pulse signal 51. This modulated optical pulse signal is input into the clock signal/feedback signal generating unit, and a clock signal 80 having a frequency f/N GHz is extracted. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a clock signal extraction method for extracting a clock signal from an optical time division multiplexed signal, and a clock signal extraction device using this method.

  As a means for increasing the number of channels that can be transmitted and received by effectively using limited communication line resources, a multiplex communication method such as time division multiplexing (TDM) has been studied. TDM multiplexes multiple channels in parallel and transmits them as time division multiplexed signals, and separates them from the time division multiplexed signals into individual channels (hereinafter “gating”) using a gate signal generated from the clock signal on the receiving side. Thus, the communication adopts a method of individually extracting and receiving information of individual channels.

  A signal that gives a time reference for a time-division multiplexed signal that is transmitted and received is called a base rate clock signal (sometimes referred to as a “clock signal” or a “reference clock signal”). The frequency of the clock signal extracted from the received signal may be equal to the bit rate indicating the number of bits transmitted / received per unit time, but may be extracted as a frequency divided by a fraction. is there. In the following description, the extracted clock signal is simply referred to as “clock signal” in any of the above cases within a range where no confusion occurs unless particularly necessary.

  When gating, a clock signal extracted from the received time division multiplexed signal is used as a gate signal at the same frequency or divided. Therefore, a technique for extracting the clock signal from the time-division multiplexed pulse signal is required, and many devices for extracting the clock signal have been studied.

  Timing jitter is mixed in the signal from the transmitter through the transmission line to the receiver. Therefore, if the clock signal is simply extracted from the received signal, this timing jitter component is included in the clock signal. For this reason, in order to accurately gate the received signal, the timing jitter is removed until it becomes sufficiently smaller than the time allocated to one bit of the signal (hereinafter, this time is sometimes referred to as “time slot”). It is necessary to extract the clock signal in the generated state.

  At present, TDM by electrical means has been realized by developing an electronic circuit that can process 40 Gbit / s signals, but it is pointed out that this level of communication speed is the limit only by electrical means. Has been. In order to realize high-speed communication exceeding this, an apparatus configured by introducing optical means in addition to electrical means is required. A clock signal extraction apparatus configured by introducing optical means in addition to electrical means may be hereinafter referred to as an optical-electric hybrid configuration clock signal extraction apparatus.

  In the optical communication system having a communication speed exceeding 40 Gbit / s as described above, an example of successful clock signal extraction has recently been reported (for example, see Non-Patent Document 1). This non-patent document 1 describes a 10 GHz clock signal from an optical pulse signal having a bit rate of 160 Gbit / s, which is the number of bits transmitted per unit time, encoded in RZ (Return to Zero). An example of successful extraction and gating by this clock signal has been reported.

  In this report, a phase-locked loop (PLL) based on a differential amplifier circuit is formed using a bidirectional electro-absorption modulator (EAM). And 10 GHz clock signal is extracted.

  The configuration of the clock signal reproduction device disclosed in Non-Patent Document 1 will be described with reference to FIG. 1 and the drawings referred to in the following description, optical signal paths such as optical fibers are indicated by thick lines, and electrical signal paths are indicated by thin lines. The numbers and symbols given to these thick and thin lines mean optical signals or electrical signals, respectively.

  The 160 Gbit / s RZ-encoded optical pulse signal 10 is branched into two by an optical demultiplexer 12 and is intensity-divided into a first optical pulse signal 14 and a second optical pulse signal 16.

  The first optical pulse signal 14 is input to the EAM 20 via the optical circulator 22. On the other hand, the second optical pulse signal 16 is added with a phase delay by the optical phase delayer 18 and input to the EAM 20 via the optical circulator 26. The first optical pulse signal 14 and the second optical pulse signal 16 input to the EAM 20 are respectively switched by the EAM 20 and input again to the optical circulators 22 and 26, respectively, and the third optical pulse signal 23 and the fourth optical pulse signal 23 are input. An optical pulse signal 27 is obtained.

  The third optical pulse signal 23 is input to the photoelectric converter 24, converted into an electric signal 25, and input to the differential amplifier 30. On the other hand, the fourth optical pulse signal 27 is input to the photoelectric converter 28, converted into an electric signal 29, and input to the differential amplifier 30. In the differential amplifier 30, a differential signal 31 between the electrical signal 25 and the electrical signal 29 is generated and output.

  The differential signal 31 is converted into a signal 33 for controlling a voltage controlled oscillator (VCO) 34 by a low-pass filter 32, and the signal 33 is input to the VCO 34. Signal 35 is extracted. The 10 GHz clock signal 35 is intensity-branched by the power divider 36 and output to the outside as a 10 GHz clock signal 35a.

  On the other hand, in order to close the PLL system, the other clock signal 35b whose intensity is branched by the power branching device 36 is multiplied by 4 by the frequency multiplier 38 and output as a 40 GHz clock signal 39. The 40 GHz clock signal 39 is input to the EAM 20 and functions as a signal for switching the first optical pulse signal 14 and the second optical pulse signal 16.

  The EAM 20 described above functions as an ultrafast electro-optic phase comparator. The first optical pulse signal 14 and the second optical pulse signal 16 are input to the EAM 20 and transmitted through the signal component. The 160 Gbit / s RZ-encoded optical pulse signal 10 and the 40 GHz driving the EAM 20 Is determined by the phase difference from the electrical signal 39.

  The electrical signal 25 and the electrical signal 29 are out of phase with each other due to the optical phase delay 18 provided between the optical demultiplexer 12 and the optical circulator 26. Due to this phase shift, a feedback signal 39 for driving the EAM 20 can be generated. The optimum value of the phase shift amount depends on the bit rate of the optical pulse signal.

  It can be seen that the PLL system of the clock signal extraction device disclosed in Non-Patent Document 1 described above is configured as follows. That is, the third optical pulse signal 23 and the fourth optical pulse signal 27 are photoelectrically converted by the photoelectric converter 24 and the photoelectric converter 28, respectively, and output. The phase of the electrical signal 25 that is the output signal is compared with the phase of the electrical signal 29 by the differential amplifier 30, and the difference signal 31 that is the difference component is removed by the low-pass filter 32 and extracted as the output signal 33. Then, by inputting the output signal 33 to the VCO 34, a PLL configured to match the phase of the oscillation signal of the VCO 34 with the phase of the 160 Gbit / s RZ-encoded optical pulse signal 10.

  However, in the above-described clock signal extraction device, when extracting a clock signal from a 160 Gbit / s RZ-encoded optical pulse signal, as is apparent from the above description, a photoelectric operation capable of a high-speed operation of 40 GHz or more is possible. Requires a converter. As the bit rate increases, that is, as the communication speed increases, the operation speed of the photoelectric converter needs to be higher. It is generally difficult to obtain a photoelectric converter that can operate at a high speed of 40 GHz or more with the current technology. In addition, it is expected that the communication speed will be higher than 160 Gbit / s in the future, and accordingly, it is necessary to obtain a photoelectric converter that operates at a higher speed.

  In order for the clock signal extraction device described above to operate as designed, the optical path length from the optical circulators 22 and 26 to the photoelectric converters 24 and 28, respectively, and the optical path from the EAM 20 to the optical circulators 22 and 26, respectively. The lengths must match exactly. In addition, the path from the optical demultiplexer 12 to the optical circulator 22, the path from the optical demultiplexer 12 to the optical phase delay 18, and the optical path length from the optical phase delay 18 to the optical circulator 26 are exactly as designed. Must be set.

If the bit rate is 160 Gbit / s, the time slot that is the time allocated to one bit of the signal is 1 / (160 × 10 ⁹ ) s = 6.3 × 10 ⁻¹² s = 6.3 ps. During this time, the distance that the light travels through the medium with a refractive index of about 1.5 (the effective refractive index of the optical fiber is about this) is (3 × 10 ⁸ m / s) × (6.3 ps) ÷ 1.5 = 12.6 × 10 ^-4 m = 1.26 mm. That is, in consideration of the structure of the connector for joining the optical element such as the optical circulator of the optical fiber, the optical path length error must be sufficiently smaller than 1.26 mm. In order to realize this configuration, a very advanced technique is required.

  Also, when operating the equipment, fluctuations in the ambient temperature of the transmission line such as optical fiber also affect the optical path length, so the fluctuation in the optical path length due to changes in the ambient temperature is sufficient compared to the above 1.26 mm The amount of change in ambient temperature must be reduced so as to decrease. That is, advanced and precise maintenance technology including ambient temperature management is required.

  Both the technology for satisfying the installation accuracy of the optical components including the optical fiber at the time of manufacturing the clock signal extraction device described above and the maintenance technology by monitoring the ambient temperature during the operation of the device are the above-mentioned non-patents. It is not feasible as the literature discloses, but it is very sophisticated.

  Therefore, the inventors of the present invention have made a prototype of a clock signal extraction device having an opto-electric hybrid configuration described below, and are researching to improve the performance (see Non-Patent Documents 2 to 4). The clock signal extraction device having an opto-electric hybrid configuration disclosed in Non-Patent Document 2 is based on the differential amplifier circuit using bidirectional EAM disclosed in Non-Patent Document 1 described above. Compared with a clock signal extraction device having a PLL system, it has the following features.

  That is, the clock signal extraction device disclosed in Non-Patent Document 2 does not require a photoelectric converter capable of high-speed operation as described above, and requires advanced technology for installing optical components such as optical fibers. It has the feature of not.

  With reference to FIG. 2, the configuration and function of the clock signal extraction device disclosed in Non-Patent Document 2 will be described. FIG. 2 is a schematic block configuration diagram of a part constituting a PLL system and a part executing optical modulation of a clock signal extraction device in a receiving device of an optical time division multiplexing communication system. A gating unit 4 that separates the time-division multiplexed signal into individual channels is also added to indicate where the clock signal output from the clock signal extraction device is used.

  The optical pulse signal 1 received by the receiving device of the optical time division multiplexing communication system is branched by the optical demultiplexer 2, and one is input to the clock signal extracting device 110 as the optical pulse signal 48 to extract the clock signal. . The other optical pulse signal 3 is input to the gating unit 4 and branched for each channel by the gating signal 80 generated based on the clock signal extracted by the clock signal extraction device 110 described above.

  Here, as an example, it is assumed that signals for four channels at a bit rate of 40 Gbit / s per channel are optically time-division multiplexed and transmitted as 160 Gbit / s RZ-encoded optical pulse signals. The configuration and operation of the clock signal extraction device 110 will be described. That is, in the gating unit 4, a 40 Gbit / s RZ-encoded optical pulse signal corresponding to 1 to 4 channels is separated and output from a 160 Gbit / s RZ-encoded multiplexed signal An explanation will be given assuming this.

  The optical pulse signal 48 branched by the optical demultiplexer 2 is input to the EAM 50. An electric signal 79 of (40−Δf) GHz is input to the EAM 50, and the optical pulse signal 48 is modulated by the electric signal 79. The electrical signal of frequency (40-Δf) GHz input to EAM 50 is a strict sine wave, but the optical signal output subjected to modulation has a pulse shape. The signal includes a periodic electrical signal as a Fourier component. Here, N is an integer of 1 or more.

  In the following explanation, when the optical signal output subjected to modulation with a strict sine wave is in the form of a pulse, the main frequency component is not referred to as including a Fourier component in a range in which confusion does not occur. In some cases, the frequency value is used as a representative value. For example, in the above case, it may be simply expressed as an electrical signal having a frequency of (40−Δf) GHz.

  The details of the operation principle of the EAM 50 will be described later. When the optical pulse signal 48 that is the input light to the EAM 50 propagates through the optical waveguide provided in the EAM 50, the absorption of the waveguide is provided. The coefficient varies according to the frequency of the electric signal 79 that is an input electric signal to the EAM 50. That is, the input light (optical pulse signal 48) propagating through the optical waveguide provided in the EAM 50 is transmitted or blocked at a frequency of (40−Δf) GHz. Here, Δf is a sufficiently small frequency value compared with 40 GHz, and is set to a value of about 250 MHz, for example.

  Hereinafter, for convenience of explanation, the EAM is generally changed to F Hz due to the phenomenon that the EAM becomes transparent or opaque at the frequency F Hz of the electric signal input to the EAM with respect to the input light. Sometimes referred to as a transmission window. That is, the above-described EAM 50 becomes transparent or opaque according to the electric signal 79 having a frequency of (40−Δf) GHz, and thus is a (40−Δf) GHz transmission window.

  When the optical pulse signal 48 is input to the EAM 50, only the component that can pass through the transmission window of (40-Δf) GHz out of the 160 Gbit / s RZ encoded optical pulse signal components is filtered. And output as a modulated light pulse signal 51. That is, a frequency obtained by inputting the optical pulse signal 48 to the EAM 50 and mixing the low frequency component Δf with a frequency 1 / N (= 1/4) of the base rate clock frequency f (= 160 GHz) of the optical pulse signal 48. Since this is a step of modulating with a modulated electric signal of (40−Δf) GHz and outputting it as a modulated optical pulse signal 51, this step may be hereinafter referred to as an optical modulation step.

  The modulated optical pulse signal 51 is input to an O / E converter 52 that converts an optical signal into an electrical signal, and is output as a first electrical signal 53. That is, since the modulated optical pulse signal 51 is input to the O / E converter 52, which is a photoelectric converter, is converted into the first electric signal 53 and output, this step is hereinafter referred to as a photoelectric conversion step. Sometimes.

  The first electric signal 53 is a frequency whose center frequency of the transmission band is four times Δf, and (4Δf) GHz among the frequency components of the first electric signal 53 by the first bandpass filter 54 of (4Δf) GHz. Are filtered out, and a second electric signal 55 of (4Δf) GHz is output. That is, the first electric signal 53 is input to the first bandpass filter 54, and only the electric signal having the frequency (NΔf) GHz of the frequency Δf multiplied by N (generally N is an integer equal to or larger than 1 and 4 in this case) is selected. Since this is a step of outputting as the second electric signal 55, this step may be hereinafter referred to as a first band pass step.

  The second electric signal 55 is input to the phase comparator 56. In the phase comparator 56, the phases of the second electric signal 55 and the third electric signal 75 having a frequency of (4Δf) GHz are compared. The third electrical signal 75 having a frequency of (4Δf) GHz is input to the quadruple multiplier 74 through the power branching device 72 from the reference signal generator 68 that outputs the electrical signal having the frequency Δf GHz. The electric signal output from the quadruple multiplier 74. If the phases of the second electrical signal 55 and the third electrical signal 75 match, the electrical signal 57 output from the phase comparator 56 is 0 V, and the electrical signal 57 is proportional to the magnitude of the phase difference. The voltage increases.

  That is, the phase comparator 56 is used to compare the phases of the second electric signal 55 having a frequency (NΔf) GHz and the third electric signal 75 having a frequency (NΔf) GHz N times the frequency Δf, and Since this is a step of outputting a difference component with respect to the phase, hereinafter this step may be referred to as a phase comparison step.

  The fourth electric signal 57 is input to the lag reed filter 58 and output as a fifth electric signal 59 having a strength averaged over time. That is, since this step is a step of averaging the difference component between the phase of the second electric signal 55 and the third electric signal 75 using the lag lead filter 58 and outputting it as a time average difference component, This step is sometimes referred to as a time average difference component output step.

  The fifth electric signal 59 is input to the VCO 60. The VCO 60 has a function of outputting a sixth electric signal 61 that is an electric signal having a frequency proportional to the voltage of the input fifth electric signal 59. For this reason, the frequency of the sixth electric signal 61 output from the VCO 60 is the same as the phase of the second electric signal 55 having a frequency (NΔf) GHz extracted from the 160 Gbit / s RZ-encoded optical pulse signal 48. The phase of the third electric signal 75 having the frequency (NΔf) GHz changes so as to match. The reason for this will be described below.

  VCO 60 sets the output frequency (center frequency of VCO 60) when the fifth electric signal 59 is 0 V to (NΔf) GHz, and the phase of the second electric signal 55 and the third electric signal When the phase of 75 matches, the sixth electric signal 61 having a frequency of (NΔf) GHz is output. That is, in order for the fourth electric signal 59 to be 0 V, the modulated optical pulse signal 51 and the electric signal 69 output from the reference signal generator 68 need to be synchronized.

  The second electric signal 55 is a signal obtained by filtering out only the frequency component of (4Δf) GHz among the frequency components of the first electric signal 53. The first electric signal 53 is converted from the modulated optical pulse signal 51 by O / E conversion. The signal is input to the device 52 and converted into an electrical signal. The modulated optical pulse signal 51 has a frequency N (40−Δf) obtained by mixing a low frequency component Δf with a frequency 1 / N (= 1/4) of the base rate clock frequency f (= 160 GHz) of the optical pulse signal 48. It is a signal modulated from GHz and output from EAM 50. Therefore, synchronizing with the phase of the second electric signal 55 means synchronizing with the phase of the electric signal having a frequency of 1 / N (= 1/4) of the frequency f (= 160 GHz) of the base rate clock of the optical pulse signal 48. It corresponds to.

  The sixth electric signal 61 output from the VCO 60 is branched by the power branching device 62, and one of them is input to the mixer 64 as a feedback signal of this PLL system. The other is output from the clock signal extraction device 110 as the extracted clock signal 80.

  Here, for convenience of explanation below, an optical pulse signal is input and modulated at a frequency obtained by mixing a low frequency component Δf with a frequency 1 / N of the base rate clock frequency f of this optical pulse signal, and the modulated optical pulse signal A portion constituted by EAM that is output as is referred to as an optical modulator. Also, this modulated optical pulse signal is inputted, and a clock signal having a frequency f / N and an electric signal having a frequency obtained by mixing a low frequency component Δf with a frequency 1 / N of the base rate clock frequency f of the optical pulse signal are provided. The portion to be output is referred to as a clock signal / feedback signal generation unit 100. Therefore, the clock signal extraction device 110 shown in FIG. 2 includes an EAM 50 that is an optical modulation unit and a clock signal / feedback signal generation unit 100.

  The mixer 64 includes a sixth electric signal 61 output from the VCO 60, and a seventh electric signal 69 that is an electric signal having a frequency (Δf) GHz output from the reference signal generator 68 via the power divider 72. Is entered. As a result, the mixer 64 outputs an eighth electric signal 65 in which a plurality of electric signal components having a frequency of (40 ± nΔf) GHz are combined.

  The eighth electric signal 65 is input to the second bandpass filter 76 whose center frequency of the transmission band is (40−Δf) GHz, and the frequency of the plurality of frequency components of the eighth electric signal 65 is (40− Only the electrical signal of Δf) GHz is filtered and output as the ninth electrical signal 77. The ninth electric signal 77 that is an electric signal having a frequency of (40−Δf) GHz is amplified by the amplifier 78, converted into the tenth electric signal 79, and input to the EAM 50. The amplifier 78 is installed as necessary to operate the EAM 50 as a transmission window, and is not necessarily installed.

  Note that the tenth electrical signal 79 is also a modulated electrical signal because it is a modulated signal input to the optical modulator.

  Here, the modulated optical pulse signal 51 output from the EAM 50 will be described. The EAM 50 is an element having a characteristic that the light transmittance is changed by applying a voltage. Therefore, an optical signal input to the EAM 50 can be modulated by applying an electric signal to the EAM 50. Although details will be described later, by increasing the amplitude of the electric signal applied to the EAM 50, the half-value width of the optical pulse constituting the modulated optical pulse signal 51 output from the EAM 50 can be reduced. In order to increase the amplitude of the sine wave, an amplifier 78 is installed, and the electric signal 77 output from the second bandpass filter 76 is amplified and input to the EAM 50 as an electric signal 79. Is the law.

  On the other hand, the electrical signal applied to the EAM 50 is strictly a sine wave, but the optical signal output from the EAM 50 is a non-sine wave. Since the optical signal output from the EAM 50 is not a sine wave but a modulated optical pulse signal 51, in addition to the repetition period component of (40-Δf) GHz, the signal is multiplied by (40-Δf) GHz. The repetition period component of is also included as a Fourier component.

  Accordingly, in the optical modulation step, the (40−Δf) GHz quadruple frequency component (160−4Δf) GHz component included in the tenth electrical signal 79 that is the control signal input to the EAM 50, and 160 Gbit / The (4Δf) GHz component that is the difference frequency from the 160 GHz component of the RZ-encoded optical pulse signal 48 of s is extracted. That is, the modulated optical pulse signal 51 output from the EAM 50 includes a (4Δf) GHz modulated component. This (4Δf) GHz component is input to the phase comparator 56 as a second electric signal 55 having a frequency of (4Δf) GHz by the OE converter 52 and the first band pass filter 54.

  On the other hand, a third electrical signal 75 having a frequency of (4Δf) GHz is also input to the phase comparator 56. In the phase comparator 56, the electric signal of the clock signal extraction device 110 is adjusted so that the phase of the second electric signal 55 having a frequency of (4Δf) GHz and the phase of the third electric signal 75 having a frequency of (4Δf) GHz are matched. The circuit components work together. As a result, the base rate clock signal of the optical pulse signal 48 that is an input signal to the clock signal extraction device 110 and the sixth electric signal 61 output from the VCO 60, that is, the clock signal extracted from the clock signal extraction device 110 A PLL circuit that synchronizes the phases of 80 is formed.

However, in the clock signal extraction device disclosed in Non-Patent Document 2 described above, since the optical modulation unit is configured using a single EAM, the width of the transmission time zone that the EAM has as a transmission window is reduced. There was a limit. For this reason, a single EAM is controlled by an electrical control signal having a frequency of (40-Δf) GHz, and the modulated optical pulse signal 51 output from this EAM includes a repetition frequency of (40-Δf) GHz. It was difficult to generate by including an optical pulse signal of (160-4Δf) GHz component, which is four times the frequency of, with sufficient intensity. As a result, there is a problem that the value of timing jitter allowed for the clock signal extraction device 110 to function as a PLL system is limited to a narrow range of values centered on zero.
"160 Gbit / s Clock Recovery With Electro-Optical PLL Using a Bidirectionally Operated Electroabsorption Modulator as Phase Comparator", C. Boerner et al., Technical Digest of OFC2003, Pape Electric F3, pp. 670-671, Vol. 2, 2003 . `` Development of 160 Gb / s clock signal extraction device '', Hiromi Tsuji, et al., 2003 IEICE Society Conference, B-10-115, 2003. "Scaleable 80 Gbit / s OTDM using a modular architecture based on EA modulators", HT Yamada et al., ECOC 2000 pp. 47-48, Munich, 2000. `` Development of 20 Gb / s × 2 optical TDM MUX / DEMUX using EA modulator '', Hiromi Yamada, et al., 1999 IEICE Society Conference, B-10-50, 1999.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a clock signal extraction method that functions as a PLL system even for timing jitter larger than the maximum timing jitter that can be achieved in phase synchronization in the conventional clock signal extraction method and clock signal extraction device. And providing a clock signal extraction device.

  In order to achieve the above object, the clock signal extraction method of the present invention comprises an optical modulation step and a clock signal / feedback signal generation step. The optical modulation step is an optical modulation step in which an optical pulse signal is input to the optical modulation unit, the optical pulse signal is modulated with a modulated electric signal, and output as a modulated optical pulse signal. This modulated electric signal is obtained by mixing an electric signal having a frequency f / N, which is a frequency 1 / N of the base rate clock frequency f of the optical pulse signal, and an electric signal having a frequency Δf. The clock signal / feedback signal generation step is a step of inputting a modulated optical pulse signal and outputting a clock signal of frequency f / N and a modulated electric signal. Then, the light modulation step is executed with the half-value width of the transmission time of the light modulation unit set to 1 / (2Nf). Here, N is an integer of 1 or more.

  As a device for realizing the above-described clock signal extraction method, the clock signal extraction device of the present invention includes an optical modulation unit and a clock signal / feedback signal generation unit. The optical modulator receives the optical pulse signal, modulates the optical pulse signal with the modulated electric signal, and outputs the modulated optical pulse signal. The modulated electrical signal is obtained by mixing an electrical signal having a frequency f / N, which is a frequency 1 / N of the base rate clock frequency f of the optical pulse signal, and an electrical signal having a frequency Δf. The clock signal / feedback signal generator receives the modulated optical pulse signal and outputs a clock signal having a frequency f / N and a modulated electric signal.

  The optical modulation unit described above is cascaded in the order of a first EAM that modulates an optical pulse signal, an optical amplifier that amplifies the optical output signal of the first EAM, and a second EAM that modulates the output of the optical amplifier. It is preferable to do this. For convenience of the following description, the optical modulation unit configured as described above is referred to as a cascade connection type optical modulation unit, and a clock signal extraction device having this cascade connection type optical modulation unit is referred to as a first clock signal extraction device.

  Further, the above-described optical modulation unit is configured by a tandem type optical modulator configured by monolithically integrating a first optical modulation region, an optical waveguide region, and a second optical modulation region that modulate an optical pulse signal in series. preferable. For convenience of the following description, a clock signal extraction device having an optical modulation unit constituted by a tandem optical modulator is referred to as a second clock signal extraction device.

  In the clock signal extraction method according to the present invention, the clock signal / feedback signal generation step described above includes a photoelectric conversion step, a first band pass step, a phase comparison step, a time average difference component output step, and a frequency voltage control. A step, a first branching step, a mixing step, a second bandpass step, an amplification step, a reference signal generation step, a second branching step, and a frequency multiplication step.

  The photoelectric conversion step is a step in which a modulated light pulse signal is input, converted into a first electric signal, and output. The first band pass step is a step of inputting the first electric signal, selecting only the electric signal component of the frequency NΔf, and outputting it as the second electric signal. In the phase comparison step, the second electric signal having the above-described frequency NΔf and the third electric signal that is an electric signal having the frequency NΔf generated by multiplying the reference electric signal having the frequency Δf generated by the reference signal generator by N times. This is a step of comparing the phases and outputting the difference component between them as a fourth electric signal.

  The time average difference component output step is a step in which the fourth electric signal output in the above-described phase comparison step is time averaged and a fifth electric signal that is a time average difference component is output. The frequency voltage control step is a step of inputting the fifth electric signal and outputting it as a sixth electric signal having a frequency f / N.

  The first branching step is a step of branching the sixth electric signal having the frequency f / N. The mixing step mixes the sixth electric signal having the frequency f / N and the seventh electric signal having the frequency Δf generated by the reference signal generator, and outputs an eighth electric signal which is a sum frequency or a difference frequency signal of both frequencies. This is a step of outputting a signal. The second band pass step inputs the eighth electric signal output in the mixing step, selects only the electric signal component having a frequency of ((f / N) -Δf), and outputs it as the ninth electric signal It is. The amplification step is a step of amplifying the ninth electric signal that is an output signal from the second bandpass filter and supplying the amplified electric signal to the optical modulation unit.

  The reference signal generation step is a step of outputting the seventh electric signal having the frequency Δf generated by the reference signal generator. The second branching step is a step of branching the seventh electric signal having the frequency Δf output from the reference signal generator. The frequency multiplying step is a step of multiplying and outputting the frequency of the seventh electric signal of the frequency Δf output from the reference signal generator. Here, N is an integer of 1 or more.

  In order to realize the above-described clock signal / feedback signal generation step, the clock signal / feedback signal generation unit includes a photoelectric converter, a first band pass filter, a phase comparator, a loop filter, a voltage controlled oscillator, And a first branching device, a mixer, a second band pass filter, an amplifier, a reference signal generator, a second branching device, and a frequency multiplier.

  The photoelectric converter receives the modulated light pulse signal, converts it into a first electrical signal, and outputs it. The first band pass filter receives the first electric signal, selects only the electric signal component of the frequency NΔf, and outputs it as the second electric signal.

  The phase comparator includes a second electric signal having the above-described frequency NΔf and a third electric signal that is an electric signal having the frequency NΔf generated by multiplying the reference electric signal having the frequency Δf generated by the reference signal generator by N. The phase is compared and the difference component between the two is output. The loop filter averages the fourth electric signal output in the above-described phase comparison step, and outputs a fifth electric signal that is a time average difference component.

  The voltage controlled oscillator receives the fifth electrical signal and outputs it as a sixth electrical signal having a frequency f / N. The first branching device branches the sixth electric signal having the frequency f / N. The mixer mixes the sixth electric signal having the frequency f / N and the seventh electric signal having the frequency Δf generated by the reference signal generator described above to obtain a first frequency or a difference frequency signal of both frequencies. 8 Outputs electrical signals. The second band pass filter filters the eighth electric signal output from the mixer and outputs a ninth electric signal that is an electric signal component having a frequency of ((f / N) −Δf). The amplifier amplifies the ninth electric signal that is an output signal from the second bandpass filter, and supplies the amplified electric signal to the optical modulation unit as a modulated electric signal. The reference signal generator described above outputs a seventh electrical signal having a frequency Δf. The second branching device branches the seventh electric signal having the frequency Δf output from the reference signal generator. The frequency multiplier multiplies the frequency of the electrical signal having the frequency Δf output from the reference signal generator and outputs the result. Here, N is an integer of 1 or more.

  According to the clock signal extraction method of the present invention comprising the optical modulation step and the clock signal / feedback signal generation step, the half-value width value of the transmission time of the optical modulation unit is set to 1 / (2Nf) in the optical modulation step. Thus, as will be described later, the frequency N ((f / N) −, which is an N-multiplied component of the frequency ((f / N) −Δf) component, included in the modulated optical pulse signal output from the optical modulation unit. The optical pulse signal component of Δf) can be maximized. That is, the optical pulse signal component of frequency NΔf, which is the difference frequency component between the optical pulse signal component of frequency N ((f / N) -Δf) and the optical pulse signal component of frequency f, is increased to the maximum to perform optical modulation. Can be output from the department. Therefore, the S / N ratio of the optical pulse signal component having the frequency NΔf can be increased, and the loop gain of the PLL can be increased. That is, the clock signal extraction method functions as a PLL system even for larger timing jitter.

  According to the first clock signal extraction device of the present invention described above, the optical modulation unit is configured by cascading two optical modulators of the first EAM and the second EAM, so that the optical modulation unit is input to the optical modulation unit. The optical pulse signal to be transmitted passes through two transmission windows, the first EAM and the second EAM. As a result, as described later, it becomes possible to narrow the half-value width of the optical pulse constituting the modulated optical pulse signal that is the output light of the optical modulation unit, and the half-value width of the transmission time of the optical modulation unit in the optical modulation step. The value can be set to 1 / (2Nf).

  According to the second clock signal extraction device of the present invention, the optical modulation unit is configured by monolithically integrating the first optical modulation region, the optical waveguide region, and the second optical modulation region that modulate the optical pulse signal in series. Therefore, the optical pulse signal input to the optical modulator passes through two transmission windows, ie, the first optical modulation region and the second optical modulation region. This makes it possible to reduce the half-value width of the optical pulse that constitutes the modulated optical pulse signal that is the output light of the optical modulator, in the same way as in the first clock signal extraction device described above. The half-value width of the transmission time of the light modulator can be set to 1 / (2Nf).

  In the clock signal extraction method described above, the clock signal / feedback signal generation step includes a photoelectric conversion step, a first band pass step, a phase comparison step, a time average difference component output step, and a time average difference component. A step of inputting and outputting as a clock signal of frequency f / N, a first branching unit step, a mixing step, a second band pass step, an amplification step, a reference signal generation step, a second branching step, And a frequency multiplication step.

  According to the clock signal / feedback signal generation step of the present invention, the clock signal and the electric modulation signal can be generated. Moreover, as will be described in detail later, in the optical modulation step, by setting the half-value width of the transmission time of the optical modulator to 1 / (2Nf), the optical pulse signal of the bit rate f and the frequency N ((f / It is possible to obtain a modulated optical pulse signal that includes a frequency component equal to fN ((f / N) −Δf) = NΔf, which is a difference frequency component from the electrical modulation signal N) −Δf). As a result, the optical pulse signal component having a frequency NΔf equal to the difference frequency component between the modulated electrical signal of frequency N ((f / N) -Δf) and the optical pulse signal of frequency f is increased to the maximum to perform optical modulation. The optical modulation pulse signal can be output from the unit. As a result, the S / N ratio of the optical modulation pulse signal component having the frequency NΔf can be increased, and the loop gain of the PLL can be increased.

  Further, the clock signal / feedback signal generation unit that realizes the above-described clock signal / feedback signal generation step includes a photoelectric converter, a first bandpass filter, a phase comparator, a loop filter, a voltage-controlled oscillator, A first branching device, a mixer, a second bandpass filter, an amplifier, a reference signal generator, a second branching device, and a frequency multiplier can be provided.

  With this configuration, the clock signal / feedback signal generation step described above can be realized, and the clock signal extraction device having a large PLL loop gain can be manufactured.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Each figure shows one configuration example according to the present invention, and only schematically shows the arrangement relationship of each component to the extent that the present invention can be understood. It is not limited to. In the following description, specific equipment and conditions may be used. However, these materials and conditions are only one of preferred examples, and are not limited to these. Moreover, in each figure, the same component is shown with the same number, and the overlapping description may be omitted.

<First embodiment>
With reference to FIG. 3, the configuration and operating principle of the first clock signal extraction device according to the first embodiment of the present invention will be described.

  The first clock signal extraction device 125 of the present invention includes an optical modulation unit 120 and a clock signal / feedback signal generation unit 100. The optical modulation unit 120 executes an optical modulation step, and the clock signal / feedback signal generation unit 100 executes a clock signal / feedback signal generation step. Since the configuration of clock signal / feedback signal generation unit 100 is the same as that in the conventional clock signal extraction device described with reference to FIG. 2, description of the configuration and operation thereof will not be repeated.

  The optical modulator 120 includes a first EAM 92 that modulates the optical pulse signal 48, an optical amplifier 94 that amplifies the first modulated optical pulse signal 93 that is an optical output signal of the first EAM 92, and the optical amplifier 94 The first EAM 92, the optical amplifier 94, and the second EAM 96 are cascaded in this order with the second EAM 96 that modulates the second modulated optical pulse signal 95, which is the output signal, and outputs it as the third modulated optical modulated pulse signal 51. Configured.

  In addition, the tenth electric signal 79 output from the amplifier 78 must be supplied to the first EAM 92 and the second EAM 96. For this reason, a phase adjuster 98 for adjusting the phase of the electric signal 88 supplied to the first EAM 92 and the second EAM 96 is provided. Details of the configuration and functions of the light modulator 120, including the functions performed by the phase adjuster 98, will be described later.

  Similarly to the case where the optical modulation unit 120 is configured using EAM as a single unit, the optical pulse signal 48 is input and 1 / N of the base rate clock frequency f of the optical pulse signal 48 is input. Of the frequency components obtained by mixing the electrical signal of the low frequency Δf with the electrical signal of the frequency f / N, only the component whose frequency is ((f / N) -Δf) is filtered by the second band pass filter 76. Modulated by the tenth electric signal 79 and output as the third modulated optical pulse signal 51.

  Since the third modulated optical pulse signal 51 is a modulated optical pulse signal obtained by inputting and modulating the optical pulse signal 48 to the optical modulator 120, the configuration and function of the optical modulator 120 will be described hereinafter. In FIG. 1, except for the case where it is necessary to distinguish between the first modulated optical pulse signal 93 and the second modulated optical pulse signal 95, in particular, instead of describing as the third modulated optical pulse signal 51, simply described as the modulated optical pulse signal 51. Sometimes.

  The clock signal / feedback signal generation unit 100 receives the third modulated optical pulse signal 51 and outputs a clock signal 80 having a frequency f / N and a tenth electrical signal 79 that is a modulated electrical signal. The tenth electric signal 79 is a difference frequency ((f) obtained by mixing an electric signal having a frequency equal to the clock signal 80 having a frequency f / N that is 1 / N of the base rate clock frequency f and an electric signal having a low frequency Δf. / N) -Δf).

  In the following description, for convenience of explanation, the base rate clock frequency of the optical pulse signal 48 is set to 160 Gbit / s, and a signal with a base rate clock frequency of 40 Gbit / s for one channel is time-multiplexed for four channels. However, of course, the technical content of the following description is not limited to this case. In addition, when it is necessary to give generality to the description, if it is necessary to display numerical values representing frequencies, etc. using alphabets, the size is about Gbit / s or GHz. In order to show this, the bit rate or frequency may not be simply expressed as f or Δf, but may be expressed with units such as (f) Gbit / s or (Δf) GHz.

  The tenth electrical signal 79 input as an electrical control signal to the optical modulation unit 120 is low in the electrical signal of 1 / N frequency (f / N) GHz of the base rate clock frequency (f) Gbit / s of the optical pulse signal. Although it is an electric signal having a frequency ((f / N) ± Δf) GHz obtained by mixing an electric signal having a frequency (Δf) GHz, its time waveform is not a strict sine wave.

  Although it is not technically realized that the time waveform of the tenth electric signal 79 input as an electric control signal to the optical modulation unit is a rectangular pulse waveform, it is assumed that the waveform is approximately a rectangular pulse waveform for the sake of brevity. An explanation will be given. Although the description will be made assuming that N = 4 in principle, the following description is valid even when N = 4.

  The modulated optical pulse signal 51 output from the optical modulation unit 120 has a frequency N of N times with ((f / 4) -Δf) GHz and ((f / 4) -Δf) GHz as fundamental frequencies as Fourier components. ((f / N) -Δf) GHz frequency component. Here, N is an integer of 1 or more. Here, paying attention to the quadruple frequency 4 ((f / 4) −Δf) GHz component corresponding to N = 4, when the duty ratio of the pulse waveform of the input tenth electric signal 79 is 1/8, The energy becomes the maximum compared to when the duty ratio is other than 1/8.

  That is, the frequency 4 ((f / 4) of the optical pulse signal 48 having a bit rate of (f) Gbit / s inputted to the optical modulation unit 120 and the tenth electric signal 79 inputted as an electric control signal to the optical modulation unit 120. ) -Δf) The frequency component equal to f-4 ((f / 4) -Δf) = 4Δf, which is the difference frequency component from the GHz component, is the pulse waveform of the tenth electric signal 79 input to the optical modulation unit. Maximum when duty ratio is 1/8.

  Incidentally, the frequency ((f / 4) -Δf) GHz component and frequency 2 ((f / 4) -Δf) GHz component corresponding to N = 1 and N = 2 of the tenth electrical signal 79 are input. When the duty ratio of the pulse waveform of the electrical signal is 1/2 and 1/4, respectively, the energy is maximized compared to other cases. That is, as the duty ratio is reduced, in other words, as the half width of the transmission window is narrower, the modulated optical pulse signal output from the optical modulation unit includes a higher frequency component.

  The above will be described in detail with reference to FIGS. 4 (A) and 4 (B). FIG. 4 (A) shows a time waveform with time on the horizontal axis. The duty ratio of the pulse waveform is a value given by d / T, the ratio of the width (time width) d of the optical pulse to the time width T occupied by one optical pulse on the time axis. In FIG. 4 (B), the duty ratio (d / T) of the pulse waveform is taken on the horizontal axis, and the amplitude ratio of the harmonic component is taken on the vertical axis, with respect to the duty ratio (d / T) of the rectangular wave. It is a figure which shows the ratio of the harmonic component contained in the square wave (refer to the catalog of Wenzel Associates published in 1995).

  According to FIG. 4B, it can be seen that, for example, when the duty ratio is d / T = 0.5, 0.25, and 0.125, the amplitudes of the first, second, and fourth harmonics are maximized, respectively. That is, for example, it is understood that the duty ratio d / T may be set to 0.125 when it is desired to output the fourth-order harmonic component to the maximum.

  Since the optical pulse signal 48 incident on the optical modulator 120 is not strictly a sine wave signal, the optical pulse signal 48 has a bit rate frequency and a Fourier component of a frequency having an integer multiple of this frequency. Yes. On the other hand, the tenth electric signal 79 input as the modulation electric signal to the light modulation unit 120 is not a sine wave signal, but has a time waveform that can be approximated by a rectangular wave having a frequency of ((f / 4) −Δf) GHz. It has a Fourier component of a frequency having a value that is an integral multiple of this frequency.

  As described above, in order to make the harmonic component of the frequency 4 ((f / 4) −Δf) GHz of the modulated optical pulse signal 51 dominant, the tenth input to the optical modulator 120. It can be said that it is most effective to set the duty ratio of the pulse waveform of the electric signal 79 to 1/8. That is, since the product of the optical pulse signal 48 having a bit rate of (f) Gbit / s and an electric signal having a frequency of 4 ((f / 4) −Δf) GHz is output in the optical modulation unit 120, as described above. And a frequency component equal to f-4 ((f / 4) −Δf) = 4Δf, which is a difference frequency component between the bit rate (f) Gbit / s and the fourth harmonic component of the tenth electric signal 79. It can be seen that in order to obtain the second electric signal 55, it is preferable to set the duty ratio of the pulse waveform of the tenth electric signal 79 input to the light modulator 120 to 1/8.

  In the above example, the duty ratio of the pulse waveform of the tenth electric signal 79 input to the optical modulation unit 120 is 1/8, that is, the time for one slot of the optical pulse signal is (1 / f) ps. The half-value width of the transmission time per cycle of the light modulation unit 120 is that the transmission time value per cycle of the light modulation unit 120 is set to (1 / (8f)) ps. This is for the case where N = 4. In general, the half-value width of the transmission time per cycle of the optical modulation unit is set to (1 / (2Nf)) ps, and the optical modulation is performed. If the step is executed, the frequency equal to fN ((f / N) −Δf) = NΔf, which is the difference frequency component between the bit rate (f) Gbit / s and the Nth harmonic component of the tenth electrical signal 79 A modulated light pulse signal 51 including the maximum component can be obtained.

  That is, in general, if the light modulation step is set to 1 / (2Nf) for the half-value width of the transmission time of the light modulator 120, the component whose frequency is N ((f / N) -Δf) is maximized. A modulated optical pulse signal 51 including a large amount can be obtained. In other words, the second electric signal 55 having the frequency NΔf, which is the difference frequency component between the modulated optical pulse signal component of the frequency N ((f / N) -Δf) and the optical pulse signal component of the frequency f, is maximized. Can be output from the first band pass filter 54.

  As a result, the S / N ratio of the second electric signal 55 having the frequency NΔf can be increased, and the loop gain of the PLL can be increased. When the optical modulation unit is configured by using only one EAM unit as employed in the conventional clock signal extraction device, the value of the half-value width of the transmission time of the optical modulation unit is set to (1 / (2Nf )) It was difficult to set to ps. In particular, as the bit rate (f) Gbit / s of the optical pulse signal input to the optical modulation unit, that is, the bit rate in communication increases, 1 / (2Nf) decreases. Therefore, as the bit rate in communication increases, it is difficult to increase the S / N ratio of the modulated optical pulse signal component 55 whose frequency output from the optical modulation unit is (NΔf) GHz.

  By establishing a technology that can narrow the half-value width of the transmission time as the transmission window of the light modulator, as described above, a clock signal extraction device that functions as a PLL system even for larger timing jitter is realized. It becomes possible.

  As described above, it is not realized that the time waveform of the tenth electric signal 79 input as the electric control signal to the light modulation unit 120 is a rectangular pulse waveform. For the sake of brevity, assuming that the waveform is approximately a rectangular pulse waveform, set the half-value width of the transmission time per cycle of the optical modulator to (1 / (2Nf)) ps, and modulate the light. It has been explained that if the step is executed, the modulated optical pulse signal 51 including the maximum frequency component equal to (NΔf) GHz can be obtained. However, although mathematical proof is omitted, even if it is assumed that the pulse waveform of the electric signal input as the electric control signal to the optical modulation unit is a pulse waveform other than the rectangular wave, the half width is reduced. It is known that the frequency component having a large multiplication order of the multiplication component having a fundamental frequency of ((f / N) −Δf) GHz of the modulated optical pulse signal output from the optical modulation unit increases. In addition, if the value of the half width of the transmission time per cycle of the optical modulation unit is set to (1 / (2Nf)) ps and the optical modulation step is executed, the frequency component equal to (NΔf) GHz It is not necessary to change the conclusion that a modulated optical pulse signal including the maximum amount can be obtained.

  Therefore, according to the first clock signal extraction device of the present invention, since the optical amplifier 94 that amplifies the first modulated optical pulse signal 93 of the first EAM 92 is installed in the optical modulation unit 120, the first modulated optical pulse signal The structure is capable of amplifying 93 light energy. In addition, the optical pulse signal 48 passes through two transmission windows, the first EAM 92 and the second EAM 96. This makes it possible to narrow the half-value width of the optical pulse that constitutes the third modulated optical pulse signal 51 that is the output light of the optical modulation unit 120, and the third modulated optical pulse signal output from the optical modulation unit 120 The half-value width of the transmission time can be set to (1 / (2Nf)) ps.

<Second Embodiment>
With reference to FIG. 5, the configuration and operating principle of the second clock signal extraction device according to the second embodiment of the present invention will be described.

  The second clock signal extraction device 135 of the present invention includes an optical modulation unit 130 and a clock signal / feedback signal generation unit 100. An optical modulation step is executed in the optical modulation unit 130, and a clock signal / feedback signal generation step is executed in the clock signal / feedback signal generation unit 100. Here, as in the first embodiment, the configuration of the clock signal / feedback signal generation unit 100 is the same as that in the conventional clock signal extraction device described with reference to FIG. The description of the operation will not be repeated.

  The optical modulator 130 includes a tandem optical modulator 112 that modulates the optical pulse signal 48 and an optical amplifier 114 that amplifies the modulated optical pulse signal 113 of the tandem optical modulator 112. As shown in FIG. 5, the optical pulse signal 48 is modulated by a tandem optical modulator 112 to become a modulated optical pulse signal 113, and this modulated optical pulse signal 113 is amplified by an optical amplifier 114 to be converted into a modulated optical pulse signal 51. The light is output from the light modulator 130.

  The optical pulse signal 48 input to the optical modulation unit 130 has a frequency component obtained by mixing an electrical signal having a low frequency Δf with an electrical signal having a frequency f / N that is 1 / N of the base rate clock frequency f of the optical pulse signal 48. Among them, only a component having a frequency of ((f / N) −Δf) is modulated by a tenth electric signal 79 obtained by filtering with a second band pass filter, and is output as a modulated optical pulse signal 51.

  The modulated optical pulse signal 51 is input to the clock signal / feedback signal generation unit 100, and a clock signal 80 having a frequency f / N and a tenth electrical signal 79, which is a modulated electrical signal, are output. Since the operation principle of the second clock signal extraction device 135 after the modulated optical pulse signal 51 is input to the clock signal / feedback signal generation unit 100 overlaps with the above description, it will not be repeated here.

  In the second clock signal extraction device 135 of the present invention, the first optical modulation region 230, the optical waveguide region 234, and the second optical modulation region that modulate the input optical pulse signal as the optical modulator constituting the optical modulation unit 130. A tandem type optical modulator 112, in which 232 is monolithically integrated in series, is employed. Thus, unlike the optical modulation unit 120 set in the first clock signal extraction device described above, the optical amplifier 94 corresponding to the optical amplifier that amplifies the optical output signal of the first EAM is not required.

  The detailed configuration and operation principle of the optical modulation unit 130 configured using the tandem optical modulator 112 will be described later. First, the configuration and operation of the second clock signal extraction device 135 of the present invention will be described. To do.

  Since the first light modulation region 230 and the second light modulation region 232 are monolithically integrated in series with the optical waveguide region 234 interposed therebetween, coupling loss occurs at the input / output ends of the first EAM and the second EAM described above. do not do. As a result, the S / N ratio of the frequency component in which the frequency of the optical modulation signal is equal to 4Δf can be further increased. As a result, the PLL loop gain in the clock signal / feedback signal generation unit 100 can be increased. This is a first improvement with respect to the first clock signal extraction device 125 described above.

  In addition, since no optical amplifier is disposed between the first light modulation region 230 and the second light modulation region 232, the first light modulation region of the electrical signal that is the modulation signal input to the second light modulation region 232 There is no need to adjust the phase of the electrical signal input to the. Therefore, the configuration of the light modulation unit is simplified. In addition, since the optical amplifier 94 is not disposed, there is no timing jitter generated from the optical amplifier 94, so that the timing jitter included in the optical modulation signal 51 can be reduced accordingly. This further reduces jitter components included in the clock signal output from the clock signal / feedback signal generation unit 100.

  Further, the first clock signal extraction device 125 and the second clock signal extraction device 135 of the present invention reduce the half-value width of the transmission time per cycle of the light modulation unit 120 or the light modulation unit 130 to 1 / (2Nf). Since the optical modulation step can be executed by setting, it is possible to obtain the modulated optical pulse signal 51 including the maximum component having the frequency N ((f / N) −Δf). As a result, the effect that the optical pulse signal component having the frequency NΔf can be maximized and the modulated optical pulse signal 51 can be output from the optical modulator 130 is common. As a result, the second electric signal 55 having a frequency of (NΔf) filtered out by the first bandpass filter 54 can be obtained effectively. That is, the S / N ratio of the second electric signal 55 having the frequency NΔf can be increased, and the loop gain of the PLL can be increased. Therefore, the first and second clock signal extraction devices of the present invention have a wide capture range (hold-in range) and hold range (hold-in range) from the viewpoint of a PLL system. It has the feature that it is excellent in.

<Electroabsorption type optical modulator>
With reference to FIG. 6, the structure and operation of an optical modulator configured using a single electroabsorption optical modulator will be described. FIG. 6 is a schematic configuration diagram of an optical modulator configured using a single electroabsorption optical modulator. The electroabsorption optical modulation element 200 includes an optical waveguide 202, and is a diode having a PN junction set at a position where the optical waveguide 202 is formed. As shown in FIG. 6, the bias power supply 210 applies a reverse potential to the diode through the coil 208 and the power branching device 242. That is, the potential of the P-side electrode 212 is set lower than the potential of the N-side electrode 214.

  The electroabsorption optical modulation element 200 is sandwiched between a first waveguide region 222 and a second waveguide region 224 to form a light absorption region 220. The reason why the first waveguide region 222 and the second waveguide region 224 are provided depends on the manufacturing convenience of the electroabsorption optical modulation element 200. That is, if the first waveguide region 222 and the second waveguide region 224 are not provided, the size of the light absorption region 220 is short, so that the total length as a diode is too short and it is difficult to mount. is there.

The dimension of the first waveguide region 222 and the dimension of the second waveguide region 224 are equal Lg ₁ , and the dimension of the light absorption region 220 is La ₁ . The dimensions of the first waveguide region 222 and the second waveguide region 224 need not be equal, but they do not necessarily have different dimensions. The value is determined by the convenience of manufacturing the electroabsorption optical modulator 200, and belongs to the design matter. On the other hand, the dimension La ₁ of the light absorption region 220 is determined by the bias voltage applied between the P-side electrode 212 and the N-side electrode 214 and the magnitude of the electrical signal. In particular, it is necessary to design in consideration of the modulation speed of the electroabsorption optical modulation element 200.

  The electroabsorption optical modulation element 200 shown in FIG. 6 is configured such that modulated light is incident as incident light from the left side as viewed in the drawing, and modulated light is output as outgoing light from the right side as viewed in the drawing. For this purpose, antireflection films 216 and 218 are respectively formed on the incident end face on the left side in the drawing and the outgoing end face on the right side in the drawing.

  The coil 208 is inserted to prevent the AC component of the electrical signal supplied from the modulated electrical signal supply unit 206 via the power branching unit 242 from being input to the bias power source 210. The capacitor 204 is inserted to prevent a bias voltage (DC voltage) supplied from the bias power source 210 from being input to the modulated electric signal supply unit 206. Incidentally, in the clock signal extraction device of the present invention, the modulated electric signal supply unit 206 corresponds to the amplifier 78, and the modulated electric signal supplied to the electroabsorption optical modulation element 200 corresponds to the tenth electric signal 79.

  In general, if no voltage is applied to the electroabsorption optical modulation element, the optical waveguide of the electroabsorption optical modulation element may be considered transparent to input light. Actually, the optical absorption coefficient of the optical waveguide of the electroabsorption optical modulation element with respect to the input light is not zero. However, for the sake of simplicity, the optical absorption coefficient is assumed to be zero here. On the other hand, when the potential on the N pole side of the electroabsorption light modulator 200 is set lower than the potential on the P pole, the light absorption coefficient of the optical waveguide of the electroabsorption light modulator 200 with respect to the input light increases as the potential difference increases. Increases and becomes opaque.

  Further, when a voltage is applied to the electroabsorption light modulation element 200 in the forward direction, the electroabsorption light modulation element 200 operates as a light emitting diode. In this case, the electroabsorption light modulation element 200 is input to the electroabsorption light modulation element 200. Since the light emitted from the light emitting diode is further added to the input light and output, it cannot be used as an optical modulator. Therefore, the bias voltage applied to the electroabsorption optical modulator 200 is the N pole of the electroabsorption optical modulator 200 even at the moment when the maximum value of the modulated electric signal is input to the electroabsorption optical modulator 200. It is necessary to set the potential on the side to be lower than the potential on the P pole side.

FIGS. 7A and 7B are diagrams for explaining the light modulation characteristics of the electroabsorption optical modulator 200, in which time is plotted on the horizontal axis and bias voltage is plotted on the vertical axis. FIG. 7 (A) shows the light modulation characteristics when the amplitude of the electric signal is (a ₁ ) V and the bias voltage is set to (−V ₁ ) V. FIG. 7B shows light modulation characteristics when the amplitude of the electric signal is (a ₂ ) V and the bias voltage is set to (−V ₂ ) V.

  In FIGS. 7A and 7B, the broken line shown as the transparent level is incident if the voltage applied to the electroabsorption optical modulation element 200 is in the range of the voltage shown by the broken line as the transparent level from 0 V. This indicates that the light can be handled as being transmitted. That is, the electroabsorption optical modulation element 200 is transparent when the applied voltage is in the range of the voltage indicated by the broken line from 0 V to the transparent level, and is applied from the voltage of the level indicated by the broken line as the transparent level. This means that when the voltage decreases (the absolute value of the applied voltage increases), it becomes opaque and functions as a transmission window.

As shown in Fig. 7 (A), when the amplitude of the electrical signal is set to (a ₁ ) V and the bias voltage is set to (-V ₁ ) V, in the time zone of the thick portion of the curve showing the electrical signal The electroabsorption light modulation element 200 is transparent. In other words, the time interval of the bold portion of the curve indicating the electrical signal that is the transparent time zone is (t ₁ ) s. On the other hand, as shown in FIG. 7 (B), when the amplitude of the electric signal is (a ₂ ) V and the bias voltage is set to (−V ₂ ) V, the portion of the curve showing the electric signal is similarly shown In this time zone, the electroabsorption optical modulation element 200 is transparent. In other words, the time interval of the thick portion of the curve indicating the electrical signal that is the transparent time zone is (t ₂ ) s.

In the description of the operation principle of the electroabsorption optical modulation element 200 as the transmission window described above, it is assumed that V ₁ <V ₂ and a ₁ <a ₂ . As is apparent from the figure, t ₁ > t ₂ . This is because the absolute value of the bias voltage is increased and the amplitude of the electrical signal is increased, so that the transmission window is transparent only in a narrower time zone centered on the maximum value taken by the waveform on the electrical signal time axis. It comes from being able to set as time. That is, by setting the absolute value of the bias voltage to a large value and increasing the amplitude of the electric signal, the width of the transmission time zone as the transmission window of the electroabsorption optical modulation element 200 can be narrowed.

  Hereinafter, for convenience of explanation, setting the absolute value of the bias voltage for the electroabsorption optical modulation element 200 to a large value may bias the electroabsorption optical modulation element 200 deeply. In addition, setting a large potential difference between the potential on the N pole side and the potential on the P pole side of the electroabsorption optical modulation element 200 may refer to applying a deep voltage to the electroabsorption optical modulation element 200. The potential difference applied to the electroabsorption optical modulation element 200 is smaller, that is, the closer to the P-pole side electric potential of the electroabsorption optical modulation element 200, the lower the potential.

  From the above description, it is apparent that the narrowing of the width of the transmission time zone as the transmission window of the electroabsorption optical modulation element 200 can be realized by deeply biasing and increasing the amplitude of the electric signal. Further, the narrowing of the width of the transmission time zone can also be realized by setting the transparent level potential shallow. This is because the potential of the transparent level can be brought close to 0 V by setting the potential of the transparent level shallow, so that the electrical signal time axis can be reduced without deep biasing and without increasing the amplitude of the electrical signal. This is because the time zone that is the transmission time zone centered on the local maximum value taken by the waveform can be set short.

In the conventional clock signal extraction device described with reference to FIG. 2, in order to narrow the half-value width of the optical pulse constituting the modulated optical pulse signal 51, the narrowing of the width of the transmission time zone as the transmission window of the optical modulation unit It is necessary to plan. The narrowing of the width of the transmission time zone as a transmission window of the EAM 50 used as the light modulation unit is realized by setting the transparent level potential shallow as described above. The potential of the transparent level can be set shallow by increasing the element length of EAM 50 (the length La ₁ of the optical waveguide in the absorption region). However, if the element length of EAM 50 is increased, the electrostatic capacity of EAM 50 can be increased. There is a limit that capacity increases and high-speed modulation becomes impossible.

  In addition to this, the power of the modulated optical pulse signal transmitted through the EAM 50 is reduced by increasing the amplitude of the electrical signal input to the EAM 50 as a control signal. Therefore, there is a problem that the power of the frequency component equal to 4Δf of the modulated optical pulse signal also decreases, and the S / N ratio for the frequency component equal to 4Δf decreases.

  In addition, the narrowing of the width of the transmission time zone as the transmission window of the EAM 50 can be realized by deeply biasing and increasing the amplitude of the modulated electric signal as described above, but an electric field that can be produced at present. There is a limit to the withstand voltage of the absorption light modulation element. For this reason, it is not always possible to realize the narrowing of the width of the transmission time zone as desired in the clock signal extraction device. For example, the withstand voltage of an electroabsorption optical modulation element that can be manufactured at present is about 5 V. In other words, it is necessary to use the electroabsorption optical modulator under conditions within this range. Further, when the amplitude of the modulated electric signal is increased, the modulated electric signal is input to the electroabsorption optical modulation element as a modulation signal, so that the temperature of the electroabsorption optical modulation element rises, and the electroabsorption optical modulation element The characteristics of the transmissive window may change and become unusable.

  That is, as in the conventional clock signal extraction device, when an electro-absorption type optical modulation element is configured by itself, an optical modulator having a desired transmission time band width can be obtained by limiting the operation speed and the withstand voltage as the optical modulator. There may be cases where narrowing cannot be realized.

  As in the conventional clock signal extraction device, when the electro-absorption type optical modulation element is configured alone to constitute the optical modulator, the above-mentioned limit exists to narrow the half-value width of the output optical pulse from the optical modulation unit, The light modulation unit of the first clock signal extraction device of the present invention comprises two electroabsorption optical modulation elements and is configured in a cascade connection, without increasing the capacitance of each electroabsorption optical modulation element, The half width of the output light pulse can be reduced. For this reason, it is possible to configure an optical modulation unit that can perform optical modulation even with a high-frequency electrical modulation signal that cannot be realized when an optical modulator is constituted by an electroabsorption optical modulation element alone.

<Cascade connection type optical modulator>
With reference to FIG. 3, the configuration and operation of cascade connection type optical modulation unit 120 will be described. The cascade connection type optical modulation unit 120 includes a first EAM 92 that modulates the optical pulse signal 48, an optical amplifier 94 that amplifies the first modulated optical pulse signal 93 that is an optical output signal of the first EAM 92, and the optical amplifier 94. A second EAM 96 that modulates a second modulated optical pulse signal 95 that is an optical output signal of the amplifier 94 is configured by cascading the first EAM 92, the optical amplifier 94, and the second EAM 96 in this order.

  Further, the power branching device 82 that branches the power of the tenth electrical signal 79, which is the modulated electrical signal output from the clock signal / feedback signal generation unit 100, into the electrical signal 84 and the electrical signal 86, and the phase of the electrical signal 86 are adjusted. A phase adjuster 98 that outputs the electric signal 88 and inputs the electric signal 88 to the second EAM 96 is provided.

  The first EAM 92 and the second EAM 96 have the same configuration as that described as the configuration of the optical modulator with reference to FIG. In the optical modulation unit 120 installed in the first clock extraction device of the present invention, the electrical signal 84 and the second EAM correspond to the modulated electrical signal supply unit 206 shown in FIG. For 96, there is an electrical signal 88.

  In order to narrow the width of the transmission time zone as the transmission window of the first EAM 92, the first EAM 92 is deeply biased and the amplitude of the electric signal 84 is set large. For this reason, the optical intensity of the first modulated optical pulse signal 93 which is the optical output signal of the first EAM 92 is significantly reduced.

  As described with reference to FIG. 7, when the amplitude of the modulated electrical signal 79 is increased and the bias voltage is set deep, a portion exceeding the transparency level of the first EAM 92 in the curve indicating the modulated electrical signal 79 ( In FIG. 7, the time band occupied by the thick line of the curve indicating the electric signal) is narrowed. Therefore, among the components of the optical pulse signal 48 incident on the first EAM 92, the component of the optical pulse signal 48 that can pass through the first EAM 92 increases as the amplitude of the electrical signal 84 is increased and the bias voltage is set deeper. , Get smaller. In practice, the optical pulse signal 48 incident on the first EAM 92 receives a loss of about 13 to 15 dB and is output as the second modulated optical pulse signal 93.

  Therefore, the second modulated optical pulse signal 93 output from the first EAM 92 is amplified by the optical amplifier 94. The optical amplifier 94 can be configured using an optical fiber amplifier or a semiconductor optical amplifier. The problem with which type of optical amplifier is used is that the optical carrier wave constituting the optical pulse signal 48 and the degree of polarization of the optical carrier wave can be controlled in the optical communication using this device. It is a problem that belongs to the device design matter that should be determined by whether or not.

  Since the optical modulation unit 120 is configured by cascading the first EAM 92, the optical amplifier 94, and the second EAM 96 in this order, the optical pulse that forms the optical pulse signal 48 passes through the transmission window of the first EAM 92. There is no guarantee that there will be an optical pulse that passes through the transmission window of the second EAM 96 at the same time as the current time. That is, the time characteristics of the transmission window of the first EAM 92 and the time characteristics of the transmission window of the second EAM 96 are matched, that is, the optical pulses constituting the optical pulse signal 48 pass through the transmission window of the first EAM 92. The phase of the modulated electrical signal applied to the first EAM 92 and the second EAM 96 is adjusted so that there is a pair of optical pulses that pass through the transmission window of the second EAM 96 at the same time. There is a need.

  A phase adjuster 98 that adjusts the phase of the modulated electric signal applied to the first EAM 92 and the second EAM 96 includes a conductive wire on the cylinder and a rod-shaped conductive wire inserted in contact with the inner wall of the cylinder. And the length of the entire conductive wire can be adjusted by sliding the rod-shaped conductive wire. The speed of propagation of light in vacuum is approximately equal to the speed of propagation of conductive lines of electrical signals. For example, the pulse interval on the spatial axis of a 40 Gbit / s optical pulse signal is approximately 7.5 mm, and that of a 160 Gbit / s optical pulse signal is approximately 1.875 mm. Therefore, the phase adjuster 98 can adjust the length of the conductive wire within a range of 7.5 mm for a 40 Gbit / s optical pulse signal and 1.875 mm for a 160 Gbit / s optical pulse signal. It can be seen that the structure is sufficient.

Next, the light modulation characteristics of the light modulation unit 120 will be described with reference to FIG. FIGS. 8A and 8B are diagrams for explaining the light modulation characteristics of the light modulation unit 120. The horizontal axis represents time, and the vertical axis represents the bias voltage. FIG. 8A shows the light modulation characteristics when the amplitude of the electric signal 84 applied to the first EAM 92 is (a ₁ ) V and the bias voltage is set to (−V ₁ ) V. FIG. 8B shows light modulation characteristics when the amplitude of the electric signal 88 applied to the second EAM 96 is (a ₁ ) V and the bias voltage is set to (−V ₁ ) V. Even if the energy of the tenth electric signal (modulated electric signal) 79 is equally branched by the power divider 82, the amplitude of the electric signal 88 is reduced by passing through the phase adjuster 98. Small compared to 84. However, since the sizes are almost the same in terms of implementation, here, in order to simplify the explanation, both are described as being equal. In other words, the electric signals applied to the first EAM 92 and the second EAM 96 are different by the amount adjusted by the phase adjuster 98, and their amplitudes are equal (a ₁ ) V.

  In the following description, it is assumed that the bias voltages applied to the first EAM 92 and the second EAM 96 are also equal. In some embodiments, this bias voltage may be different, but the effect of this bias voltage determines the transparent time interval of the transmission window as described above. Similarly, it is necessary to set the bias voltage in consideration of the withstand voltage of the electroabsorption type optical modulation elements constituting the first EAM 92 and the second EAM 96.

  In FIGS. 8A and 8B, the broken line shown as the transparency level has the same meaning as described in FIGS. 7A and 7B. That is, if the voltage applied to the first EAM 92 and the second EAM 96 is within a range of voltage from 0 V to a level indicated by a broken line as a transparent level, the optical pulse signal 48 and the second modulated light that are the respective incident lights This indicates that the pulse signal 95 is within a range that can be treated as being transmitted.

As shown in FIG. 8 (A), when the amplitude of the electric signal 84 applied to the first EAM 92 is (a ₁ ) V and the bias voltage is set to (−V ₁ ) V, the electric signal 84 is shown. The first EAM 92 is transparent in the time zone of the thick part of the curve. That is, the time interval of the transparent time zone is (t ₁ ) s.

  The optical pulse that constitutes the first modulated optical pulse signal 93 that is transmitted through the first EAM 92 and output is a curve that represents the electrical signal 84 shown in FIG. It is an optical pulse component that has been able to pass through the time zone shown in bold. That is, the optical pulse that constitutes the first modulated optical pulse signal 93 is a product component of the transmission pulse of the first EAM 92 and the optical pulse that constitutes the optical pulse signal 48. Compared with the pulse, the half width as the optical pulse is narrow.

A signal obtained by amplifying the first modulated optical pulse signal 93 by the optical amplifier 94 is a second modulated optical pulse signal 95. Therefore, the intensity of the second modulated optical pulse signal 95 is the same as that of the optical pulse signal 48, but the half width is narrow. Therefore, the optical signal 51 that is modulated and output by the second EAM 96 passes through a transmission window that is modulated by an electrical modulation signal having a narrowed half-value width of the time waveform as shown in FIG. Can be thought of as equivalent to That is, it can be considered that the time interval of the transparent time zone is (t ₂ ′) s.

  That is, the optical pulse transmitted through the second EAM 96 includes the transmission waveform of the second EAM 96, which is a transmission window modulated by the electrical modulation signal whose half width of the time waveform is narrowed, and the second modulated optical pulse signal 95. Since it can be regarded as a product component of the optical pulse constituting the optical pulse, the half width as the optical pulse is narrower than the optical pulse constituting the first modulated optical pulse signal 93.

For the single electroabsorption optical modulator that constitutes the optical modulator described with reference to FIGS. 7A and 7B, the amplitude of the modulated electric signal is increased and the bias voltage is set deeper. The width of the optical pulse of the modulated optical pulse signal obtained by the above is (t ₂ ) s. On the other hand, the width of the optical pulse of the modulated optical pulse signal obtained by the cascaded optical modulation unit described above is (t ₂ ′) s, which is apparent when comparing FIG. 7 (B) and FIG. 8 (B). As shown, t ₂ > t ₂ ′.

  That is, by using the optical modulation unit 120 configured by cascade connection of the first EAM 92, the optical amplifier 94, and the second EAM 96 in this order, the optical modulation unit is configured using a single electroabsorption optical modulation element. Compared with the case where the optical modulator is used, the width of the transmission time zone as the transmission window of the light modulation unit 120 can be equivalently reduced.

<Tandem type optical modulator>
The structure and operation of the tandem optical modulator will be described with reference to FIG. FIG. 9 is a schematic configuration diagram of a tandem optical modulator. The tandem light modulation element 240 includes a light guide 236, and is a diode having a PN junction set at a position where the light guide 236 is formed. As shown in FIG. 9, the bias power supply 210 applies a reverse potential as a diode through the coil 208, similar to the electroabsorption optical modulator 200 shown in FIG. That is, the potentials of the P-side electrodes 212a and 212b are set lower than the potential of the N-side electrode 214. The potentials of the P-side electrodes 212a and 212b may be set to be different from each other. However, this is merely a design matter, and in the following description, both potentials are set equal to each other for the sake of brevity. To do.

  The tandem type light modulation element 240 is sandwiched between the first light absorption region 230 and the second light absorption region 232, and a waveguide region 234 is formed. Unlike the electroabsorption optical modulation element 200 described above, portions corresponding to the first waveguide region 222 and the second waveguide region 224 are not provided. This is because, in the tandem type light modulation element 240, the first light absorption region 230 and the second light absorption region 232 are provided, so that the entire length of the element can be secured to a length sufficient for manufacturing the device.

The dimensions of the first light absorption region 230 and the second light absorption region 232 are equal to La ₂ , and the dimension of the waveguide region 234 is Lg ₂ . The size of the first light absorption region 230 and the size of the second light absorption region 232 do not have to be the same, but there is no necessity of making them different from each other. This value is determined by the convenience of manufacturing the tandem type light modulation element 240 and belongs to the design matter. On the other hand, the dimension Lg ₂ of the waveguide region 234 is determined by the bias voltage applied between the P-side electrodes 212a and 212b and the N-side electrode 214 and the magnitude of the electric signal. In particular, it is necessary to design in consideration of the modulation speed of the tandem light modulation element 240.

  The tandem light modulation element 240 shown in FIG. 9 is configured such that modulated light is incident as incident light from the left side as viewed in the drawing, and modulated light is output as outgoing light from the right side as viewed in the drawing. For this reason, the antireflection films 226 and 228 are formed on the incident end face on the left side in the drawing and the outgoing end face on the right side in the drawing, respectively.

  The roles of the coil 208, the modulated electric signal supply unit 206, the bias power supply 210, and the capacitor 204 are the same as described with reference to FIG. The modulated electric signal supply unit 206 corresponds to the amplifier 78 shown in FIG. 5, and the modulated electric signal supplied to the tandem type optical modulation element 240 corresponds to the tenth electric signal 79. Further, the power branching device 140 that branches the power of the tenth electric signal 79 shown in FIG. 5 corresponds to the power branching device 140 shown in FIG.

In the embodiment of the present invention, the size of the first light absorption region 230 and the size La ₂ of the second light absorption region 232 are set to be equal to the light absorption region 220 of the electroabsorption light modulation element 200. On the other hand, the dimension Lg ₂ of the waveguide region 234 is set to be shorter than the sum 2 Lg ₁ of the dimensions of the first and second optical waveguide regions of the electroabsorption optical modulator 200. Accordingly, the total length 2La ₂ + Lg ₂ of the tandem type light modulation element 240 was set to be shorter than twice the total length 2Lg ₁ + La ₁ of the electroabsorption light modulation element 200. Since the dimension Lg ₂ of the waveguide region 234 can be set as short as possible in the range where electrical insulation between the P-side electrodes 212a and 212b is possible, the total length 2Lg ₁ + La of the electroabsorption optical modulation element 200 It is possible to set the length to be 1.5 times shorter than ₁ .

Further, the main reason why the shorter the dimension Lg ₂ of the optical waveguide region 234 is, the _second reason is that the optical pulse constituting the optical pulse signal 48 is the second at the same time as the time when it passes through the transmission window of the first light absorption region 230. This is because it is necessary to make it possible to assume that there is a pair of light pulses that pass through the transmission window of the light absorption region 232 as well. That is, as compared with the time slot of the optical pulse signal 48, the time for light to propagate through the waveguide region 234 needs to be sufficiently short and negligible.

If the dimension La ₁ of the light absorption region 220 is increased in the electroabsorption optical modulator 200, the light absorption coefficient can be increased without increasing the absolute value of the voltage applied between the N-side electrode 214 and the P-side electrode 212. However, the capacitance of the electroabsorption optical modulation element 200 increases, and it becomes difficult to follow a modulated electric signal having a high frequency. That is, the frequency characteristics of the electroabsorption optical modulator 200 are deteriorated. In order to avoid this problem, the first clock signal extraction device of the present invention employs a cascade connection type optical modulation unit configured to cascade connect electroabsorption optical modulation elements in two stages.

  The cascade connection type optical modulation unit 120 includes a part that makes the first modulated optical pulse signal 93 incident on the optical amplifier 94 and does not occur in the structure of the optical modulator configured using a single electroabsorption optical modulation element, and There is a loss in the portion where the second modulated optical pulse signal 95 is incident on the second EAM 96. In these portions, the light emitted from the optical waveguide of the electroabsorption optical modulator is input to the core of the optical fiber, or conversely, the light emitted from the optical fiber core is input to the optical waveguide of the electroabsorption optical modulator. Or make it incident. Here, a lens is used to improve the coupling efficiency, but there are losses due to reflection on the lens surface, the difference between the propagation mode of the optical waveguide of the electroabsorption optical modulator and the propagation mode of the optical fiber, etc. A loss based on

  The loss resulting from these does not occur in the tandem light modulation element 240 in which the waveguide region 234 is monolithically formed between the first light absorption region 230 and the second light absorption region 232. Since a common optical waveguide is monolithically connected at the boundary between the first light absorption region 230 and the waveguide region 234 and at the boundary between the waveguide region 234 and the second light absorption region 232, optical loss occurs in this portion. It is because it does not.

  Regarding signal loss, a superior clock signal extraction apparatus can be configured by using the tandem type optical modulator 112 as compared with the cascade connection type optical modulation unit 120. However, on the other hand, in the cascade connection type optical modulation unit 120, the first modulated optical pulse signal 93 output from the first EAM 92 is amplified by the optical amplifier 94 and then input to the second EAM 96. . Therefore, there is a feature that the time interval of the transparent time zone as the light transmission window of the second EAM 96 is easily reduced equivalently.

  In the tandem light modulation element 240, the optical pulse signal modulated in the first light absorption region 230 propagates through the waveguide region 234 without being amplified and is input to the second light absorption region 232. Therefore, at the stage of input to the second light absorption region 232, the intensity of the optical pulse signal modulated in the first light absorption region 230 is the second modulation input to the second EAM 96 in the cascade connection type optical modulation unit 120. Small compared to optical pulse signal intensity. For this reason, it is difficult to equivalently reduce the time interval of the transparent time zone as the light transmission window of the second light absorption region 232.

  However, in the tandem type light modulation element 240, since the light modulation region is divided into the first light absorption region 230 and the second light absorption region 232, the parasitic capacitances of the first and second modulation regions, respectively. An effect equivalent to applying a deep bias voltage substantially can be obtained without increasing. That is, the following effects can be obtained as compared with the case where the light modulation step is performed using the field effect type light modulation element in which the light modulation region is provided at only one place.

  When performing a light modulation step using a field effect type light modulation element provided in only one light modulation region, a field effect type light modulation element is used to reduce the half-value width of the transmission time of the light modulation unit. It is necessary to adopt either a method of applying a bias voltage deeply or setting a long light modulation part of the field effect type light modulation element. Even if the bias voltage is applied deeply, the bias voltage is limited by the withstand voltage of the field effect type optical modulation element. Moreover, if the light modulation part of the field effect light modulation element is set long, the parasitic capacitance increases, and it becomes difficult to input a high-frequency modulated electric signal. Therefore, by using the tandem type light modulation element 240, it is possible to relax the restriction due to the withstand voltage of the field effect type light modulation element and the increase in parasitic capacitance.

  Whether to adopt the configuration of the first clock signal extraction device 125 or the second clock signal extraction device 135 comprehensively considers the bit rate of the optical pulse signal for extracting the clock signal, the intensity of the optical pulse signal, etc. It is a matter to be determined.

  Further, in the tandem light modulation element 240 described above, two light absorption regions are provided. However, a structure in which three or more light absorption regions are provided may be realized. Also, in the configuration of the cascade connection type optical modulation unit described above, a configuration using two EAMs is used, but a configuration using three or more units can of course be used. However, increasing the number of light absorption regions or the number of EAMs used increases the complexity of the device itself, and causes problems such as new timing jitter from the inserted optical amplifier. These points are also matters to be determined in consideration of the bit rate of the optical pulse signal for extracting the clock signal, the intensity of the optical pulse signal, and the like in the same manner as described above.

It is a schematic block diagram of a conventional clock signal extraction device. It is a schematic block diagram of a conventional clock signal extraction device. 1 is a schematic block configuration diagram of a clock signal extraction device of a first exemplary embodiment. It is a figure which shows the ratio of the harmonic component contained in the rectangular wave with respect to the duty ratio (d / T) of a rectangular wave. FIG. 5 is a schematic block configuration diagram of a clock signal extraction device of a second exemplary embodiment. It is a schematic block diagram of the light modulation part comprised using EAM. It is a figure where it uses for description of the optical modulation characteristic of EAM. It is a figure where it uses for description of the optical modulation characteristic of the cascade connection type | mold optical modulation part of this invention. It is a schematic block diagram of a tandem type optical modulator.

Explanation of symbols

2: Optical demultiplexer
4: Gating part
12: Optical demultiplexer
18: Optical phase retarder
20, 50, 200: Electro-absorption modulator (EAM)
22, 26: Optical circulator
24, 28: Photoelectric converter
30: Differential amplifier
32: Low-pass filter
34: Voltage controlled oscillator
36, 62, 72, 82, 140, 242: Power divider
38: Frequency multiplier
52: O / E converter
54: First bandpass filter
56: Phase comparator
58: Lug reed filter
60: VCO
64: Mixer
68: Reference signal generator
74: Quadruple multiplier
76: Second bandpass filter
78: Amplifier
92: 1st EAM
94, 114: Optical amplifier
96: Second EAM
98: Phase adjuster
100: Clock signal / feedback signal generator
112: Tandem type optical modulator
120, 130: Light modulator
200: Electroabsorption optical modulator
202, 236: Optical waveguide
204: Condenser
206: Modulated electrical signal supply unit
208: Coil
210: Bias power supply
212, 212a, 212b: P-side electrode
214: N side electrode
216, 218, 226, 228: Antireflection film
220: Light absorption area
222: First waveguide region
224: Second waveguide region
230: 1st light absorption area
232: Second light absorption region
234: Waveguide region
240: Tandem light modulator

Claims (5)

A modulated electric signal obtained by inputting an optical pulse signal to the optical modulator and mixing an electric signal having a frequency f / N of 1 / N of the base rate clock frequency f of the optical pulse signal with an electric signal having a frequency Δf An optical modulation step of modulating the optical pulse signal and outputting it as a modulated optical pulse signal,
The modulated optical pulse signal is input, and a clock signal / feedback signal generating step for outputting a clock signal of frequency f / N and the modulated electric signal is provided,
A clock signal extraction method, wherein the light modulation step is executed by setting a half-value width of a transmission time of the light modulation unit to 1 / (2Nf).
Here, N is an integer of 1 or more. The clock signal extraction method according to claim 1,
The clock signal / feedback signal generation step includes
The photoelectric conversion step of inputting the modulated light pulse signal, converting it to a first electric signal and outputting it,
A first band pass step for inputting the first electric signal, selecting only an electric signal component of the frequency NΔf and outputting it as a second electric signal;
The phase of the second electric signal having the frequency NΔf and the phase of the third electric signal that is the electric signal having the frequency NΔf generated by multiplying the reference electric signal having the frequency Δf generated by the reference signal generator by N times A phase comparison step for outputting the difference component between the two as a fourth electrical signal;
A time average difference component output step of averaging the fourth electric signal output in the phase comparison step and outputting a fifth electric signal that is a time average difference component;
A frequency voltage control step for inputting the fifth electrical signal and outputting it as a sixth electrical signal of frequency f / N;
A first branching step for branching the sixth electrical signal of frequency f / N;
The sixth electric signal having the frequency f / N and the seventh electric signal having the frequency Δf generated by the reference signal generator are mixed to obtain an eighth electric signal which is a sum frequency or difference frequency signal of both frequencies. A mixing step that outputs
A second bandpass step of inputting the eighth electrical signal output in the mixing step, selecting only an electrical signal component having a frequency of ((f / N) -Δf) and outputting it as a ninth electrical signal;
An amplification step of amplifying the ninth electric signal, which is an output signal from the second bandpass filter, and outputting the modulated electric signal;
A reference signal generating step for outputting a seventh electric signal having the frequency Δf;
A second branching step of branching the seventh electric signal of frequency Δf output from the reference signal generator;
A frequency multiplication step of multiplying and outputting the frequency of the seventh electric signal of the frequency Δf output from the reference signal generator. By inputting an optical pulse signal, the modulated electric signal obtained by mixing an electric signal having a frequency f / N of 1 / N of a base rate clock frequency f of the optical pulse signal and an electric signal having a frequency component Δf, An optical modulator that modulates the optical pulse signal and outputs the modulated optical pulse signal; and
Inputting the modulated optical pulse signal, comprising a clock signal of frequency f / N and a clock signal / feedback signal generator for outputting the modulated electrical signal,
The optical modulator includes a first electroabsorption optical modulator that modulates the optical pulse signal, an optical amplifier that amplifies an optical output signal of the first electroabsorption optical modulator, and an output of the optical amplifier And a second electro-absorption optical modulator that is cascade-connected in this order.
Here, N is an integer of 1 or more. By inputting an optical pulse signal, the modulated electric signal obtained by mixing an electric signal having a frequency f / N of 1 / N of a base rate clock frequency f of the optical pulse signal and an electric signal having a frequency component Δf, An optical modulator that modulates the optical pulse signal and outputs the modulated optical pulse signal; and
Inputting the modulated optical pulse signal, comprising a clock signal of frequency f / N and a clock signal / feedback signal generator for outputting the modulated electrical signal,
The optical modulation unit includes a tandem optical modulator configured such that a first optical modulation region, an optical waveguide region, and a second optical modulation region that modulate the optical pulse signal are monolithically integrated in series. A clock signal extraction device.
Here, N is an integer of 1 or more. The clock signal extraction device according to claim 3 or 4,
The clock signal / feedback signal generation unit
A photoelectric converter that inputs the modulated light pulse signal, converts it into a first electric signal, and outputs it,
A first band-pass filter that inputs the first electric signal, selects only an electric signal component of the frequency NΔf, and outputs it as a second electric signal;
The phase of the second electric signal having the frequency NΔf and the phase of the third electric signal that is the electric signal having the frequency NΔf generated by multiplying the reference electric signal having the frequency Δf generated by the reference signal generator by N times A phase comparator that outputs the difference component between the two as a fourth electrical signal;
A loop filter that time-averages the fourth electrical signal output in the phase comparator and outputs a fifth electrical signal that is a time-averaged difference component;
A voltage-controlled oscillator that inputs the fifth electrical signal and outputs it as a sixth electrical signal of frequency f / N;
A first branching device for branching the sixth electric signal of frequency f / N;
The sixth electric signal having the frequency f / N and the seventh electric signal having the frequency Δf generated by the reference signal generator are mixed to obtain an eighth electric signal which is a sum frequency or a difference frequency signal of both frequencies. A mixer that outputs
A second band-pass filter that filters the eighth electrical signal output from the mixer and outputs a ninth electrical signal that is an electrical signal component having a frequency of ((f / N) −Δf);
An amplifier that amplifies a ninth electrical signal that is an output signal from the second bandpass filter, and supplies the amplified electrical signal to the optical modulator as a modulated electrical signal;
A reference signal generator for outputting the seventh electric signal having a frequency Δf;
A second branching device for branching the seventh electrical signal having a frequency Δf output from the reference signal generator;
A clock signal extraction device comprising: a frequency multiplier that multiplies and outputs the frequency of the electric signal having the frequency Δf output from the reference signal generator.

Images